| Characteristic | Overall N = 1801 |
ABC N = 301 |
ACB N = 301 |
BAC N = 301 |
BCA N = 301 |
CAB N = 301 |
CBA N = 301 |
p-value2 |
|---|---|---|---|---|---|---|---|---|
| Age (years) | 32 (8) [18, 45] | 33 (8) [19, 44] | 31 (7) [18, 42] | 31 (7) [20, 44] | 33 (8) [19, 45] | 32 (7) [21, 44] | 31 (7) [20, 44] | 0.6 |
| Race | 0.9 | |||||||
| Â Â Â Â Black | 64 (36%) | 12 (40%) | 13 (43%) | 12 (40%) | 10 (33%) | 8 (27%) | 9 (30%) | |
| Â Â Â Â White | 52 (29%) | 8 (27%) | 9 (30%) | 7 (23%) | 7 (23%) | 12 (40%) | 9 (30%) | |
| Â Â Â Â Other | 64 (36%) | 10 (33%) | 8 (27%) | 11 (37%) | 13 (43%) | 10 (33%) | 12 (40%) | |
| Female | 81 (45%) | 13 (43%) | 11 (37%) | 16 (53%) | 14 (47%) | 12 (40%) | 15 (50%) | 0.8 |
| 1 Mean (SD) [Min, Max]; n (%) | ||||||||
| 2 Kruskal-Wallis rank sum test; Pearson’s Chi-squared test | ||||||||
No apparent difference in starting value by period or treatment, so looking at change from baseline in PK
Carryover effects and period were excluded for parsimony, as the likelihood ratio test indicated that they did not significantly improve model fit
Notation: \(i =\) subject, \(j =\) treatment,
\[ \begin{align} Y_{ij} &= \mu + \tau_1\text{I(Gel B)} + \tau_2\text{I(Gel C)} + b_{i} + \varepsilon_{ij},\\ b_{ij} &\sim N(0, \sigma_i^2), \\ \varepsilon_{ij} &\sim N(0, \sigma^2), \end{align} \]
Findings: Significantly less decrease in gels vs pill A
## Model df AIC BIC logLik Test L.Ratio p-value
## model 1 5 9718.129 9743.052 -4854.064
## model_large 2 8 9721.615 9761.493 -4852.807 1 vs 2 2.513838 0.4728
## Linear mixed-effects model fit by maximum likelihood
## Data: baseline_long
## AIC BIC logLik
## 9718.129 9743.052 -4854.064
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 11.76905 19.69016
##
## Fixed effects: bloodVL_change ~ treatment
## Value Std.Error DF t-value p-value
## (Intercept) 121.16111 1.360734 898 89.04097 0
## treatmentGel B -10.88889 1.469661 898 -7.40912 0
## treatmentGel C -11.65556 1.469661 898 -7.93078 0
## Correlation:
## (Intr) trtmGB
## treatmentGel B -0.54
## treatmentGel C -0.54 0.50
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -3.40327599 -0.57002844 0.04432071 0.61132413 2.48072577
##
## Number of Observations: 1080
## Number of Groups: 180
Findings: Significantly less decrease in gels vs pill A
## Model df AIC BIC logLik Test L.Ratio p-value
## model 1 5 10773.34 10798.26 -5381.670
## model_large 2 8 10778.04 10817.92 -5381.022 1 vs 2 1.295485 0.7302
## Linear mixed-effects model fit by maximum likelihood
## Data: baseline_long
## AIC BIC logLik
## 10773.34 10798.26 -5381.67
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 15.67261 32.86382
##
## Fixed effects: skinVL_change ~ treatment
## Value Std.Error DF t-value p-value
## (Intercept) 205.52778 2.092094 898 98.24021 0
## treatmentGel B -32.40556 2.452934 898 -13.21094 0
## treatmentGel C -41.92222 2.452934 898 -17.09065 0
## Correlation:
## (Intr) trtmGB
## treatmentGel B -0.586
## treatmentGel C -0.586 0.500
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -2.99812316 -0.62341504 0.04451491 0.66568981 2.81739908
##
## Number of Observations: 1080
## Number of Groups: 180
Carryover effects and period were excluded for parsimony, as the likelihood ratio test indicated that they did not significantly improve model fit
Q: include treatment X adherence and treatment X AE interaction terms???
Skin/Blood PK ~ Random Intercept (Subject) + Treatment + Adherence + Carryover + Period
## Model df AIC BIC logLik Test L.Ratio p-value
## model 1 7 5307.160 5337.201 -2646.580
## model_large2 2 11 5290.549 5337.757 -2634.275 1 vs 2 24.61057 0.0001
## Linear mixed-effects model fit by maximum likelihood
## Data: merged
## AIC BIC logLik
## 5307.16 5337.201 -2646.58
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 0.003649564 32.52911
##
## Fixed effects: skinVL_change ~ treatment + overall_adhere + overall_safety
## Value Std.Error DF t-value p-value
## (Intercept) 162.22787 7.582825 356 21.394119 0
## treatmentGel B -24.41464 3.625130 356 -6.734832 0
## treatmentGel C -36.99593 3.543320 356 -10.441036 0
## overall_adhere 1.81506 0.274724 356 6.606870 0
## overall_safety1 -55.77639 5.201704 356 -10.722715 0
## Correlation:
## (Intr) trtmGB trtmGC ovrll_
## treatmentGel B -0.509
## treatmentGel C -0.443 0.534
## overall_adhere -0.945 0.311 0.234
## overall_safety1 0.139 -0.079 -0.032 -0.205
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -3.43058828 -0.67881147 0.02915247 0.64404884 3.53027310
##
## Number of Observations: 540
## Number of Groups: 180
## Linear mixed-effects model fit by maximum likelihood
## Data: merged
## AIC BIC logLik
## 5290.549 5337.757 -2634.275
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 0.004174974 31.79624
##
## Fixed effects: skinVL_change ~ treatment * overall_adhere + treatment * overall_safety
## Value Std.Error DF t-value p-value
## (Intercept) 238.04771 21.667665 352 10.986311 0.0000
## treatmentGel B -108.95724 23.377772 352 -4.660719 0.0000
## treatmentGel C -122.25653 23.950925 352 -5.104460 0.0000
## overall_adhere -1.13762 0.818852 352 -1.389291 0.1656
## overall_safety1 -29.61261 8.743642 352 -3.386759 0.0008
## treatmentGel B:overall_adhere 3.40490 0.904914 352 3.762682 0.0002
## treatmentGel C:overall_adhere 3.39045 0.924719 352 3.666464 0.0003
## treatmentGel B:overall_safety1 -40.84761 12.155471 352 -3.360430 0.0009
## treatmentGel C:overall_safety1 -37.04463 12.856933 352 -2.881296 0.0042
## Correlation:
## (Intr) trtmGB trtmGC ovrll_ ovrl_1 trGB:_ trGC:_
## treatmentGel B -0.927
## treatmentGel C -0.905 0.838
## overall_adhere -0.993 0.921 0.899
## overall_safety1 0.114 -0.106 -0.103 -0.148
## treatmentGel B:overall_adhere 0.899 -0.986 -0.813 -0.905 0.134
## treatmentGel C:overall_adhere 0.880 -0.815 -0.988 -0.886 0.131 0.801
## treatmentGel B:overall_safety1 -0.082 0.117 0.074 0.107 -0.719 -0.171 -0.094
## treatmentGel C:overall_safety1 -0.078 0.072 0.111 0.101 -0.680 -0.091 -0.158
## tGB:_1
## treatmentGel B
## treatmentGel C
## overall_adhere
## overall_safety1
## treatmentGel B:overall_adhere
## treatmentGel C:overall_adhere
## treatmentGel B:overall_safety1
## treatmentGel C:overall_safety1 0.489
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -3.405995796 -0.615618058 0.002664942 0.610769689 3.605667548
##
## Number of Observations: 540
## Number of Groups: 180
## Model df AIC BIC logLik Test L.Ratio p-value
## model 1 7 4860.258 4890.299 -2423.129
## model_large2 2 11 4851.039 4898.246 -2414.519 1 vs 2 17.21954 0.0018
## Linear mixed-effects model fit by maximum likelihood
## Data: merged
## AIC BIC logLik
## 4860.258 4890.299 -2423.129
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 4.74164 21.00209
##
## Fixed effects: bloodVL_change ~ treatment + overall_adhere + overall_safety
## Value Std.Error DF t-value p-value
## (Intercept) 98.73209 5.172962 356 19.086184 0.0000
## treatmentGel B -6.81051 2.354117 356 -2.893020 0.0041
## treatmentGel C -9.10003 2.295483 356 -3.964320 0.0001
## overall_adhere 0.93216 0.187884 356 4.961396 0.0000
## overall_safety1 -26.34806 3.460017 356 -7.615009 0.0000
## Correlation:
## (Intr) trtmGB trtmGC ovrll_
## treatmentGel B -0.513
## treatmentGel C -0.443 0.538
## overall_adhere -0.949 0.327 0.247
## overall_safety1 0.131 -0.079 -0.032 -0.195
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -3.5553528690 -0.5673073777 0.0005900677 0.6043037231 3.0550971969
##
## Number of Observations: 540
## Number of Groups: 180
## Linear mixed-effects model fit by maximum likelihood
## Data: merged
## AIC BIC logLik
## 4851.039 4898.246 -2414.519
##
## Random effects:
## Formula: ~1 | ptid
## (Intercept) Residual
## StdDev: 5.176441 20.56066
##
## Fixed effects: bloodVL_change ~ treatment * overall_adhere + treatment * overall_safety
## Value Std.Error DF t-value p-value
## (Intercept) 131.74236 14.431943 352 9.128525 0.0000
## treatmentGel B -39.19101 15.315595 352 -2.558895 0.0109
## treatmentGel C -52.11513 15.650359 352 -3.329964 0.0010
## overall_adhere -0.36462 0.545375 352 -0.668563 0.5042
## overall_safety1 -11.39117 5.811353 352 -1.960158 0.0508
## treatmentGel B:overall_adhere 1.31391 0.591905 352 2.219795 0.0271
## treatmentGel C:overall_adhere 1.73440 0.603583 352 2.873511 0.0043
## treatmentGel B:overall_safety1 -25.67859 8.049524 352 -3.190076 0.0016
## treatmentGel C:overall_safety1 -17.98165 8.518057 352 -2.111003 0.0355
## Correlation:
## (Intr) trtmGB trtmGC ovrll_ ovrl_1 trGB:_ trGC:_
## treatmentGel B -0.924
## treatmentGel C -0.901 0.841
## overall_adhere -0.993 0.919 0.895
## overall_safety1 0.113 -0.109 -0.107 -0.147
## treatmentGel B:overall_adhere 0.895 -0.986 -0.816 -0.901 0.138
## treatmentGel C:overall_adhere 0.875 -0.818 -0.988 -0.881 0.135 0.804
## treatmentGel B:overall_safety1 -0.080 0.118 0.075 0.104 -0.717 -0.173 -0.096
## treatmentGel C:overall_safety1 -0.075 0.073 0.113 0.098 -0.678 -0.093 -0.160
## tGB:_1
## treatmentGel B
## treatmentGel C
## overall_adhere
## overall_safety1
## treatmentGel B:overall_adhere
## treatmentGel C:overall_adhere
## treatmentGel B:overall_safety1
## treatmentGel C:overall_safety1 0.492
##
## Standardized Within-Group Residuals:
## Min Q1 Med Q3 Max
## -3.63070078 -0.59111860 -0.02603496 0.59219448 2.76927257
##
## Number of Observations: 540
## Number of Groups: 180